To: djgpp AT delorie DOT com
Subject: Re: timing instructions [code, LONG]
Message-ID: <19970210.173501.4967.1.chambersb@juno.com>
References: <5dlh98$qrl AT gummy DOT cs DOT indiana DOT edu>
From: chambersb AT juno DOT com (Benjamin D Chambers)
Date: Mon, 10 Feb 1997 20:32:51 EST


On 9 Feb 1997 16:58:00 -0500 jestandi AT cs DOT indiana DOT edu (Jeff Standish)
writes:
>Now, my question is, if you look at the two sets of results below, 
>you'll
>see than they are almost identical, _except_ for integer +/-, which
>literally takes twice as long to run when Win95 is installed.  I'm
>baffled by the result.  I don't think it has anything to do with
>cache misses, since that will typically only hurt the first loop.
>This is essentially the same code I use for benchmarking different
>machines (except for the code to get ellapsed time), and I've never 
>seen
>the likes of this before.  What about win95 would cause only integer 
>+/-
>to take twice as long to execute?
I would assume that nothing in the background is using the FPU, so it can
run unimpeded, but that the integer unit is being used in the background,
which would cause interuptions in the piping (or something...)

...Chambers